At the Keystone Symposium, Neurodegenerative Diseases: New Molecular Mechanisms on Wednesday, debate on the last plenary talk continued around the dinner table and well into the poster session. Marc Tessier-Lavigne, Genentech Inc., San Francisco, California, described a novel role for APP in neurodegeneration, a role that appears completely independent of Aβ toxicity and that might finally explain why only certain neurons bear the brunt of Alzheimer disease pathology. Tessier-Lavigne reported that the extracellular domain of APP serves as a ligand for death receptor 6 (DR6), an orphan member of the tumor necrosis factor receptor superfamily. He showed that when the N-terminus of APP (N-APP) binds DR6, it sets off an apoptotic cascade in embryonic spinal neurons that targets both axons and cell bodies. The work indicates that N-APP has a role in axonal pruning and neural cell death in development, but it also raises the possibility that similar events occur in mature neurons in the brain. “Finding that an N-terminal fragment of APP is a ligand for DR6 came as a complete surprise, and finding that APP is involved in a self-destruction mechanism like this immediately suggested that perhaps it could contribute to Alzheimer’s disease,” said Tessier-Lavigne in an interview with this reporter after the talk. The study was coincidentally published in Wednesday’s Nature.

“This is very intriguing. I think the story is very convincing because there is a lot of very complementary data, biochemical, cell biological, etc., and I do not doubt that it is relevant for APP biology,” said Bart De Strooper, University of Leuven, Belgium, who was at the meeting. He was not involved in the study. “Of course, when you have this type of data I think it is fair to ask how relevant it is for Alzheimer’s disease. I think it is too early to say if it is equivalent to the amyloid hypothesis, for example, because there is no human or clinical data,” he cautioned.

Tessier-Lavigne and colleagues identified the N-APP/DR6 interaction when studying neuronal development. The embryo is a tough environment for new neurons with many more being formed than typically needed. Those that do not make proper connections are eventually weeded out by a process that involves axonal pruning and cell death. First author Anatoly Nikolaev and colleagues found that DR6 plays a key role in this process, kicking off an apoptotic pathway in commissural, motor, and sensory neurons that depends on activation of caspase 6 in axons and caspase 3 in the soma only. Nikolaev and colleagues mimicked this axon degeneration by withdrawing trophic support (such as nerve growth factor, or NGF) from neurons in culture, but if they blocked DR6 at the same time—with an antibody, by knocking it down with RNAi, or by genetic knockout—they were able to prevent axonopathy and cell death. Collaborating with Todd McLaughlin and Dennis O’Leary at the Salk Institute, they also showed that DR6 regulates neuron death and axonal pruning not just in cell culture, but also in vivo in a mouse model.

APP entered the picture when Tessier-Lavigne and colleagues considered what might activate DR6. The protein is a cell surface receptor with no known ligands. In fact, the researchers were not even sure if a ligand was necessary for its activation, but when they incubated neurons with a soluble DR6 ectodomain construct, it prevented degeneration following trophic factor withdrawal. That suggested that a ligand, mopped up by the soluble DR6 protein, was necessary for the process. In support of this, the researchers found that the DR6-AP—the DR6 ectodomain bound to alkaline phosphatase to act as a reporter—detected proteins (100 kDa and 35 kDa) in conditioned medium following NGF withdrawal suggesting that those proteins may be the DR6 ligands that promote axon degeneration.

The researchers took a leap of faith when they decided to immediately focus on APP as a potential ligand. Tessier-Lavigne explained that in many respects APP fit the bill, as it is shed from the cell surface, it is tied to neurodegeneration already through AD, and it is highly expressed in developing neurons. In initial experiments, the researchers found that DR6-AP bound to APP on the surface of COS-1 cells. A polyclonal antibody to the N-terminal of APP also recognized the same 100 and 35 kDa proteins that turn up in conditioned medium after trophic deprivation. An antibody to the C-terminal of sAPPPβ, which is released upon BACE cleavage, also detected the 100 kDa protein in the conditioned medium, and also a 55 kDa protein. The antibody binding patterns suggest that after trophic withdrawal, APP is cleaved by BACE to yield the ~100 kDa sAPPβ, which is further cleaved, by an unknown protease, to yield the 55 and 35 kDa fragments. Tessier-Lavigne said that it is not clear whether this second cleavage is necessary for activation of the DR6 pathway.

A range of additional experiments supports the idea that APP sets axons off on a suicidal slippery slope. Degeneration of sensory neurons by trophic withdrawal is blocked by antibodies to N-APP and by knocking down APP by RNAi. BACE inhibitors also blocked the degeneration in cultured neurons, but it could be restored by adding a purified N-terminal fragment of APP (amino acids 1-286). Finally, the affinity of the N-terminus of APP for DR6 is very high (EC50 for binding is around 4.5 nM) and the interaction seems fairly specific since the researchers found that N-APP only reacts with one of the nine other members of the TNF receptor superfamily they tested, p75, and at 60-fold lower affinity.

Do these data suggest an entirely new model for neurodegeneration in AD? “What it does is it opens up additional possibilities for how APP could be involved in the pathological process and therefore opens up new potential therapeutic targets as well,” said Tessier-Lavigne in a post-talk interview. The data could, for example, explain why specific neurons are targeted in the disease despite widespread expression of APP in the CNS. Tessier-Lavigne showed that DR6 expression in the mature brain is enriched in sites of known AD pathology, including the hippocampus and forebrain cholinergic neurons, suggesting that under certain conditions, shedding of the APP ectodomain might trigger a self-destruct pathway that contributes to neurodegeneration in those very same neurons.

There are other facets of the disease that this model might have more difficulty explaining, however. David Holtzman, Washington University, St. Louis, Missouri, noted during questions that there are familial AD mutations that occur in the Aβ region of APP and that it would be hard to reconcile how they could fit into this model. Tessier-Lavigne agreed that he needs to determine whether and how such mutations might tie in to his mechanism. He also pointed out that the model is not mutually exclusive of the amyloid hypothesis, and could even be complementary. His lab is in the process of crossing DR6 knockout mice with AD mouse models to see what impact the death receptor pathway has on pathology in those systems.

In a post-talk interview, Chris Link, University of Colorado, Boulder, also raised the issue of tau pathology and the formation of neurofibrillary tangles (NFTs). “I think in some people’s view, including mine, Alzheimer’s is ultimately a tauopathy, and it is not clear how this model leads to tau hyperphosphorylation,” he said. Tessier-Lavigne suggested that caspase 6 activation, which occurs downstream of N-APP/DR6 binding, might tie APP to tau pathology. He brought up that work from Andrea LeBlanc’s lab, for example, shows that caspase 6 and tau fragments generated by the protease are elevated in aging and in mild cognitive impairment (see Albrecht et al., 2007). “It is possible that [caspase 6 activation] could lead to tau aggregation because there is good evidence that tau fragments play a role, and phosphorylation could be secondary to fragmentation,” said Link. “The issue is that although you see NFTs, and certainly tau hyperphosphorylation in neurons, they still look like neurons. If they really had degenerated then they would be gone.”

Also, in a News & Views written for Nature, Donald Nicholson, Merck Research Laboratories, Rahway, New Jersey, noted that caspase 6 can also cleave APP in almost the same location as BACE, raising the possibility that the caspase feeds back to generate more N-APP, thereby amplifying the apoptotic process or spreading it to other cells. “Such potential secondary effects are hard to ignore, particularly because they might be relevant to Alzheimer’s disease,” writes Nicholson.

Obviously, there is much more work to do to corroborate these results and to explore the mechanism in greater detail. In addition to mouse model work, Tessier-Lavigne said he wants to look for evidence of activation of the DR6 pathway in the adult human brain and particularly in the AD brain. “Since we have additional players in a biochemical pathway, we would like to know if there are any mutations associated with them that represent risk factors in Alzheimer’s disease,” he added. In fact, DR6 is located very close to a potential susceptibility locus on chromosome 6. “At the same time we are trying to develop potential therapeutic candidates to interfere with steps in this pathway and to use them in preclinical models. If that looks promising, we’d obviously consider moving toward the clinic.”—Tom Fagan.

Nikolaev A, McLaughlin T, O'Leary DD, Tessier-Lavigne M. APP binds DR6 to trigger axon pruning and neuron death via distinct caspases. Nature. 2009 Feb 19;457(7232):981-9. Abstract


Make a Comment

To make a comment you must login or register.

Comments on this content

  1. The paper from Marc Tessier-Lavigne and colleagues provides compelling evidence that DR6 mediates axonal pruning and degeneration induced by trophic-factor withdrawal in developing neurons, and that the extracellular sequences of APP bind to DR6 with high affinity and specificity. The identification of DR6 as the receptor for secreted APP and the link of APP to axonal degeneration through the DR6/caspase 6 pathway are extremely exciting. However, its physiological significance needs to be further established both in developing neurons and in adult brain. The findings that APP/DR6 activation requires BACE, but not α-secretase cleavage, that BACE processing is followed by other cleavage events, and that this pathway is Aβ independent are intriguing and leave open many interesting questions. Although the study addresses a developmental function, it could offer novel insights to AD pathogenesis and therapeutic intervention as APP/DR6/caspase 6 are expressed in both developing neurons and adult brains.

  2. It's interesting that an androgen receptor coactivator interacts with DR6 and that androgen induces the expression of APP.


    . Androgen receptor coregulator ARA267-alpha interacts with death receptor-6 revealed by the yeast two-hybrid. Sci China C Life Sci. 2004 Oct;47(5):442-8. PubMed.

    . Amyloid precursor protein is a primary androgen target gene that promotes prostate cancer growth. Cancer Res. 2009 Jan 1;69(1):137-42. PubMed.

  3. Reading this remarkable paper from the Tessier-Lavigne group feels like drinking water from a fire hose (six figures in the main text + 18 multipanel, supplementary figures). This study unambiguously shows that DR6 is involved in axonal pruning and neuronal degeneration with caspase 3 and 6 as the downstream players. The readers of this forum might be wondering how this study fits in with what we know about APP function and whether the proposed ligand (~35 kDa N-terminal fragment) is relevant to AD pathogenesis.

    1. It seems likely that APP performs different functions in the peripheral nervous system (PNS) as opposed to the central nervous system (CNS). This study is focused on PNS neurons and the uncanny similarity in neuromuscular abnormality seen in DR6-/- mice (Figure 6) and APP/APLP2 double-KO mice (1) is consistent with APP regulating axonal elongation/pruning in the PNS. However, in the CNS the story seems to be different. The lack of APP and APLP2 (and APLP1) results in abnormal migration of neurons (2), a finding supported by other studies (3) and interaction with Fe65 as a possible underlying mechanism (4) for the observed migration phenotype. Nonetheless, this can’t be the complete story since APP is also expressed in non-neuronal tissues where axonal pruning or neuronal migration is irrelevant. Obviously, more surprises await us.

    2. The relevance of the ~35 kDa N-terminal fragment of APP to AD pathology, an exciting idea as it is to all Aβ skeptics including me (5), is at best tenuous. One, these observations are made in the PNS neurons and one doesn’t know whether this will also hold true for the CNS neurons. Second, there is no evidence that the ~35 kDa fragment observed in trophic factor deprived medium and which binds DR6-AP (Figure 4) is actually toxic to neurons; all the subsequent toxicity experiments performed in the paper used recombinant APP1-286-His protein. Indeed, the crucial piece of data—IP using DR6-Fc followed by Western blot using APP-Nt antibody—is not shown in the paper.

    3. One piece of information that cannot be easily reconciled with the current findings is that more than 90 percent of APP is cleaved by the α-secretase (at least in the CNS neurons) and far too many studies have shown the sAPP-α to be neurotrophic. This group finds little or no sAPP-α in their culture medium. Perhaps PNS neurons behave differently from the CNS neurons.

    4. Finally, it is interesting that DAPT treatment also protected neurons from anti-NGF induced degeneration (Suppl. Figure 17). Inhibition of γ-cleavage should block generation and release of AICD, which is known to induce apoptosis. Could AICD be an important player, too?

    In any case, these are truly exciting and important findings that will, undoubtedly, be repeated in other labs and will further stimulate studies focused on APP function. Even if some of the details might change as more groups repeat these observations, this study provides a potential mechanism for the action of APP in the PNS during embryonic development. That is a solid step forward.


    . Defective neuromuscular synapses in mice lacking amyloid precursor protein (APP) and APP-Like protein 2. J Neurosci. 2005 Feb 2;25(5):1219-25. PubMed.

    . Cortical dysplasia resembling human type 2 lissencephaly in mice lacking all three APP family members. EMBO J. 2004 Oct 13;23(20):4106-15. PubMed.

    . A critical function for beta-amyloid precursor protein in neuronal migration revealed by in utero RNA interference. J Neurosci. 2007 Dec 26;27(52):14459-69. PubMed.

    . Essential roles for the FE65 amyloid precursor protein-interacting proteins in brain development. EMBO J. 2006 Jan 25;25(2):420-31. PubMed.

    . Reassessing the amyloid cascade hypothesis of Alzheimer's disease. Int J Biochem Cell Biol. 2009 Jun;41(6):1261-8. PubMed.

  4. This is a compelling study that provides a biological “raison d’être” for an N-terminal secreted APP fragment in axonal pruning. Its interaction with the DR6 receptor is triggered by trophic factor deprivation in cultured motor, commissural and sensory neurons. A related role for the DR6 receptor is also evident in vivo in DR6 deficient mice since axons of the phrenic nerve overshoot the motor endplate at the developing neuromuscular junction of the diaphragm muscle. Interestingly, axonal sprouting was previously reported in the diaphragm muscle of P0 APP/APLP2 double knockout mice (Wang et al., 2005). Since the β-secretase derived ectodomain of APLP2 is also a ligand for DR6, failure of DR6 signaling during development of the neuromuscular junction in the APP/APLP2 DKO may be operative. The demonstrated involvement of BACE activity in this process suggests that BACE1 and 2 knockout mice might also display this phenotype.

    sAPPβ and the N-terminal 35 kDa APP fragment are generated by NGF withdrawal and bind the DR6 receptor to activate axonal pruning. Furthermore, APP(1-286) (~35 kDa) is shown to be sufficient for this effect. Since α-secretase acts at the cell surface, this fragment could just as readily be produced by cleavage of sAPPα. Is BACE activity predominant in neurons under these conditions and how does DR6-mediated axonal pruning occur without cell body death during development? Is there local recruitment of β-secretase activity to targeted axons? The observation that an Aβ antibody (Aβ33-42) capable of blocking Aβ1-42 toxicity is unable to reverse axonal pruning resulting from NGF deprivation suggests that Aβ does not mediate axonal pruning. However, incubation of γ-secretase inhibitors partly blocked axonal degeneration induced by NGF deprivation suggesting that other γ-secretase derived products contribute to axonal degeneration in this paradigm. Alternatively, γ-secretase inhibition may lead to a redistribution of surface APP away from the axon surface. Could this be sufficient to produce partial rescue of axonal degeneration mediated by trophic deprivation? Abundant DR6 expression is observed in the adult mouse hippocampus and cortex raising the possibility that this death receptor contributes to the selective vulnerability of these regions in AD. Is there a sufficient amount of APP on the surface of adult nerve axons, and if not, is there a physiological event that can trigger surface APP expression and this pathway in Alzheimer disease?


    . Defective neuromuscular synapses in mice lacking amyloid precursor protein (APP) and APP-Like protein 2. J Neurosci. 2005 Feb 2;25(5):1219-25. PubMed.

  5. I recommend the primary paper.

  6. The data reported in this exciting paper can potentially open new ways to interpret AD pathogenesis at the molecular level. Of course, much more data are needed to assess the relevance for AD of the APP-DR6 interaction; however, when confirmed, such a new pathway may not necessarily be alternative to the amyloid hypothesis. In my opinion, one out of the many points that will need consideration is determining DR6 location on the cell membrane. This could link the APP-DR6 pathway with the effect of cell membrane cholesterol on AD pathogenesis, providing further clues on the cholesterol-AD relation. For example, it could be of interest to study the effects, if any, on the APP-DR6 pathway of increasing or reducing membrane cholesterol, and the resulting modifications of lipid raft stability. The finding that APP/DR6 activation requires BACE, but not α-secretase cleavage, adds further points to be addressed in the light of a possible link with membrane cholesterol content, raft stability, and behavior.

Comments on Primary Papers for this Article

  1. This paper provides extensive and compelling evidence for β-APPs (shed from the cell surface by β-secretase) being a major agonist for the cell death receptor DR6, a key signaling pathway in the neuronal pruning that takes place during development. Finally, a clear study on the normal biological function of APP: a seminal paper, in my view. However, too much is made of possible implications to AD. For instance, this mechanism does not explain any of the FAD mutations in either APP, PS1, or PS2 (except perhaps adding an additional mechanism for the pathogenesis of the APP Swedish mutation near the β-secretase cleavage site).

    View all comments by Michael Wolfe


Paper Citations

  1. . Activation of caspase-6 in aging and mild cognitive impairment. Am J Pathol. 2007 Apr;170(4):1200-9. PubMed.
  2. . APP binds DR6 to trigger axon pruning and neuron death via distinct caspases. Nature. 2009 Feb 19;457(7232):981-9. PubMed.

Further Reading


  1. . APP binds DR6 to trigger axon pruning and neuron death via distinct caspases. Nature. 2009 Feb 19;457(7232):981-9. PubMed.

Primary Papers

  1. . APP binds DR6 to trigger axon pruning and neuron death via distinct caspases. Nature. 2009 Feb 19;457(7232):981-9. PubMed.